[00:24] rick_h_, you're killing me bro http://imgur.com/oe41jad [00:24] jcastro_: yea, I want to get those pulled out [00:25] jcastro_: they were demo bundles [00:25] should I file this? or you got it? [00:25] also, we don't have a bug wrt. the URLs being complex, should I file that too? [00:25] I've got it. I'll ask bac to remove them tomorrow [00:26] urls will get fixed in upcoming project. I've already got a document plotting out every url we'll support [00:26] and they'll get cleaned up some with machine view, spent time today planning that out [00:26] or you mean the bundle urls themselves? [00:27] yeah [00:27] Makyo: you get it squared away? That would make sense as revisionless urls was a recent addition [00:27] Makyo: and it doesn't happen on all bundles [00:27] jcastro_: yea which? [00:27] lp:~/charms/bundles//bundle [00:27] makes me cry [00:27] oh, yea well that's by design. Not much I can do with that atm though there's plans for that as well [00:28] And I am going to solve that by making deploying bundles from github be magical. [00:28] hah [00:28] well we'll have real bundle ingestion at some point [00:28] and that'll mean they're all in the cs: or something [00:28] juju quickstart https://github.com/castrojo/mongodb-bundle [00:28] is what you want to give me [00:28] I am certain. :) [00:29] heh, that's too much magic. We can't plan out how people lay out their github repo file structure [00:29] yeah, so if it's not right, error out [00:29] we'll constantly be getting "but I have my bundle in a folder by name" [00:29] or "I've got 3 bundles in there" [00:29] but that's ok [00:39] rick_h_, Want to discuss, also includes schemaless URLs. Will be tomorrow's task either way, just need more direction tmrw morning [00:39] Makyo: ok, ping me when you get in then and we'll chat [00:39] * rick_h_ checks the calendar [00:40] Makyo: yea, I've got nothing before the standup tomorrow [00:40] rick_h_, sure. If nothing else, shoot me an email. Details are on the card. Basically, getting charm urls of "mysql", curious if we want to infer cs: and -LATEST [00:40] well charmworld does that [00:40] charmworld figures out the right thing to load...hmmm but that's in prooof [00:41] Sure, but cs is charmworld. Just want to make sure the right direction. [00:41] Makyo: ok, will look and we can chat tomorrow [00:41] rick_h_, Excellent, thanks! [00:42] things will change wrt to cs btw.. [00:43] have a look at casey's doc to the list [00:43] hazmat: yep === rogpeppe1 is now known as rogpeppe === rogpeppe1 is now known as rogpeppe [12:38] morning rick_h_ [12:39] bac: morning [12:40] rick_h_: i saw your conversation with jorge last night about test bundles [12:40] rick_h_: unfortunately we don't have a tool for removing bundles [12:40] oh? we can't remove the branches and they go away? [12:40] whoops [12:41] rick_h_: nope, ingest doesn't remove things that disappear [12:41] we need 'ingest --autoremove' :) [12:41] bac: ok, bummer [12:41] rick_h_: a while back benji added a script for removing charms [12:41] bac: and I assume that renaming them won't help us? [12:42] because they're pulled in based on that name? [12:42] rick_h_: i've deleted my bundle branch from LP and it is still in mjc [12:42] ok [12:43] i think we need to write a script analogous to benji's remove_charm.py [12:43] bac: ok, yea. benji added a remove from charm tool and that exists. Maybe we should look to add a mirror for bundles. [12:43] i think we agree! [12:43] bac: I'll add a card in the backlog for that then. Maybe try to get at it in the next 2week iteration so we can garden bundles before the march 12th push [12:43] rick_h_: we also need to figure out which ones to can [12:44] right [12:44] well, we'll be working with ecosystems to get some that are promulgated, and I'd say any others owned by us that we don't need can go. [12:44] * bac bristles a bit at jcastro's characterization.... [12:45] yea, it's a little bit unfair [12:45] but he's making his point. Now that we show that, those are at the top of the visible results which is a bit painful [12:46] yep, they need to go. of course, we just discovered a display bug by having those complicated ones prominently listed [12:46] yep, exactly [12:46] every feature adds another feature :) [14:10] frankban: when you get into the api calls layout, can you create cards for other calls we need to support vs the create/destory machine and container? [14:44] rick_h_ did you see my comments on the PR? [14:45] hatch: yep, and I'm fine with them. Gave it a +1 [14:45] coolio [14:45] sometimes review comments end up "did you think about?" or "I don't obviously see" [14:47] :) [15:02] rick_h_: I see only two machine related calls: AddMachines and DestroyMachines, and those AFAICT also handle containers. The other changes involve adding ToMachineSpec to ServiceDeploy and AddServiceUnits. I'll create a card for that if it does not take place in my current branch. How does it sound? [15:04] frankban: cool sounds good [15:18] of course within two hours of the telco guy coming to give me a new modem the internet starts working properly [15:18] :/ [15:18] "I SWEAR it was broken" [15:18] lol [15:18] trouble maker [15:18] it's not often I hope that the internet will break [15:18] haha [15:23] rick_h_ any word on your kernel patch? [15:24] hatch: just "try to use latest kernel" and I've not had time to mess with it [15:25] ahh ok - I'm waiting to hear about a success from you before I go for Trusty on this thing...I'd really like to get out of OSX land [15:25] :) [15:27] oh cool chrome devtools now shows the ::before elements in the inspector [15:30] YUS [15:30] there goes the internet [15:30] 3s latency for the win! [15:31] you get a fix? [15:31] * hatch goes to get his 14,400 baud modem [15:31] lol, no it's broken again [15:31] the guy is coming sometime after 11 so I need it to stay broken until then [15:31] lol [15:32] oh, 3s lately, 3ms [15:32] * rick_h_ added an m in there to make it nice [15:33] haha :) [15:33] local comp store has one of those fancy asus ones on order for me but I figure I might as well get a new modem too [15:44] rick_h_, ping, around if you want to chat. [15:44] Makyo: sure thing [15:45] Makyo: invite in bound [15:46] https://plus.google.com/hangouts/_/72cpi3mcfnema37uu7cu52iv44?hl=en [15:46] Makyo: ^ [15:48] rick_h_ here is the local upgrade UI now.... https://www.evernote.com/shard/s219/sh/90cdf63b-9a99-46ed-876f-9217247ae46e/dbdbb49f65ef5ffedfe94fe28a9d5b78 [15:48] ugh the internet is working agin [15:50] jujugui call in 10 [15:50] lol hatch [15:50] haha [15:51] oh well at least I saved those ping gists [15:51] if the guy even knows what a ping is.... [15:52] (the guy on the phone didn't know what latency was) [15:56] hmm I think fakebackend fails when deploying the same local charm twice [15:56] yup [15:58] jujugui call in 2 [15:59] https://bugs.launchpad.net/juju-gui/+bug/1285223 [15:59] <_mup_> Bug #1285223: fakebackend error when deploying the same local charm twice [16:23] Makyo, rick_h_: the changes i made to support revision-less bundles were all done in the deployer. no changes were made to charmworld. [16:24] bac: ok right. I don't think this is anything to do with revisions to be honest. [16:24] Makyo, rick_h_: that work did add a method 'get_qualified_charm_url' to the deployer library that can be used to do what needs to be done [16:24] bac: it's more of the charmworld finding charms for the bundle input [16:25] bac: right, but the gui needs charm data to get the charm metadata like icons from charmworld. [16:25] rick_h_: well, it involves transforming a minimally-specified charm id to a fully-qualified, non-revisioned one, right? [16:25] this is all pre deployer file [16:25] bac: right, and making sure that charmworld has that to return to the gui as metadata in the bundle json responses [16:26] rick_h_: sure. i'd prefer we do it at the time the json is created rather than modifying the id stored in the bundle at ingest time so that it continues to work after the default series changes [16:27] bac: well the default series is defined in the bundle so we can write that out [16:27] bac: and if the bundle updates it we reingest and it gets updated [16:27] ah, true [16:29] bac: but you're right that the goal is that the gui requests for information on a bundle needs to include the metadata for the charms and right now it's broken. [16:29] bac: regardless of technique used, I just want to avoid adding the charm finding logic to the gui in JS when we've got the work in charmworld atm [16:29] rick_h_: sure [16:35] ccccccbtujivdbknginddhvickuujhddijlnrhhvdikb [16:37] Same. [16:39] heh, curses [16:41] lol [17:27] rick_h_: I suppose that with the machine related stuff we will only support core >= 1.18, correct? [17:28] frankban: yes, I would think so. Was it api driven and stable back in 1.16? [17:29] rick_h_: it seems 1.16 does not include add/destroy machines [17:29] frankban: then yea, definitely. [17:31] rick_h_: cool, I guess these are the first "plural" API calls, i.e. they can be used to add/remove multiple machines/containers in bulk. I suspect this will help us in our "commit all the things" story, so cool [17:32] yay! I'm a little worried about how we're going to end up having to do that. Love that they're getting that multi call stuff in [17:36] rick_h_: looking at the backlog: I suspect all the add/destroy machine/coontainer API cards will go away with my current branch (just two ne go.js calls to rule them all) [17:36] frankban: ok, I moved them to on deck but cool to clear out [17:37] frankban: I'm still getting my head around trying to plan all the parts small enough for cards. We tend to create so many on the fly [17:37] it'll be interesting on friday [17:38] rick_h_: yeah, I've found to be quite hard some time to subdivide tasks before at least a bit of prototyping. it's not so bad to create new cards on the fly IMHO [17:38] back [17:39] frankban: yea, I've got to keep room for that in our 2 week planning. So it seems like not that many cards but I know it'll be tough to get it all done in two weeks. [17:39] learning experience :) [17:39] hatch: welcome to the land of the living and sub-second ping times? [17:40] not sure I lost all of my ip's lol [17:40] so now I need to figure out what the ip of the router is [17:40] trace your way out to google and see when you hit something 192. super looking [17:41] 2-3ms times on wifi [17:41] awww yeah [17:42] odly though internally i'm on a 172 ip [17:42] guess I'll need to change that [17:43] that's kind of odd that they would use that [18:35] has anyone played with flask before? there's a presentation tonight on it and i'm going since it's the first technical talk in english i've been able to attend. [18:36] bac, yeah, I love it. [18:37] bac: yep, it's interseting. It's lighter pyramid based on armin's stuff [18:37] cool [18:37] I don't get a lot of benifit over pyramid so I've not kept it around for anything just tinkered [18:38] i went to some lightening talks in spanish. understood more than i thought i would. [18:38] very cool === abentley__ is now known as abentley [19:46] annnnnd there goes the internet again === hatch_ is now known as hatch [22:02] * Makyo gets Charmworld mostly working. Must walk dogs, then keep at it. [22:27] charmworld doesn't have a sweet makefile like we do? [22:44] Oh, it totally has a sweet makefile. [22:44] make run just dies. [22:44] lol [22:44] super sweet [22:51] rick_h_ I'd like to move the bug in High to Urgent for frankban to do in the morning as it's blocking me now